Freshwater crabs ( Decapoda : Brachyura : Pseudothelphusidae , Trichodactylidae ) from the state of Roraima , Brazil : species composition , distribution and new records

The species composition and distribution of the pseudothelphusid and trichodactylid freshwater crabs from the state of Roraima, Brazil, are updated based mainly on documented records deposited in the collections of the Instituto Nacional de Pesquisas da Amazônia (Manaus) and the Museu Integrado de Roraima (Boa Vista). This paper provides records for the occurrence of ten species (six Pseudothelphusidae, four Trichodactylidae). Three species are recorded from Roraima for the first time: the pseudothelphusids Fredius beccarii (Coifmann, 1939) and Fredius fittkaui (Bott, 1967), and the trichodactylid Moreirocarcinus laevifrons (Moreira, Nauplius, 26: e2018011 All content of the journal, except where identified, is licensed under a Creative Commons attribution-type BY. orcid.org/0000-0002-2118-6153 orcid.org/0000-0003-2426-8936 orcid.org/0000-0003-4858-2575 original arTiCle This article is part of the special series offered by the Brazilian Crustacean Society in honor to Ludwig Buckup in recognition of his dedication and contributions to the development of Carcinology CORRESPONDING AUTHOR Célio Magalhães celiomag@inpa.gov.br SUBMITTED 02 November 2017 ACCEPTED 29 March 2018 PUBLISHED 04 June 2018 Guest Editors Alessandra Angélica de Pádua Bueno and Sandro Santos DOI 10.1590/2358-2936e2018011

Carcinological surveys made during the last decade in several areas throughout the state revealed that its freshwater crab fauna is more diverse and widespread than currently known.The purpose of this paper is to document the occurrence and distribution of several species of both pseudothelphusid and trichodactylid crabs in Roraima, and to extend the range of distribution of some of these species into the Amazon River basin in Brazil.

Study area
Roraima is the northernmost of the Brazilian states, located between the geographic coordinates 05º16'N and 01º25'S, and 058º55'W and 64º48'W, and bordered by Venezuela in the west and north, and Guyana in the east (Ferreira et al., 2007).The hydrographic system is dominated by the Branco River, whose watershed covers 187,540 km 2 , comprising 83% of the state territory (Carvalho, 2015).The river is formed by the confluence of the Uraricoera River and Tacutu River and flows northeast-southwest for 566 km.Other main rivers are the Xeruini and Jauaperi Rivers (tributaries of the Negro River), and Jatapu River (tributary of the Amazon River), all located in the south and southeastern portion of the state (Santos et al., 1985;Ferreira et al., 2007;Carvalho, 2015).Roraima is situated within the boundaries of the Guyana Shield region as defined in Huber (1995) and in the Remarks and Discussion sections we use the hydrologic limits of the region as defined in Lujan and Armbruster (2011: 212, fig. 13.1).

Data collection
The specimens were obtained either from field surveys, conducted in several localities throughout the state between April 2014 and April 2015, or from material deposited in the collections of the Instituto Freshwater crabs from Roraima, Brazil Nauplius, 26: 2018011 Nacional de Pesquisas da Amazônia, Manaus (INPA), Instituto Venezolano de Investigaciones Científicas, Caracas (IVIC), Museu Integrado de Roraima, Boa Vista (MIRR), Museu Nacional, Universidade Federal do Rio de Janeiro, Rio de Janeiro (MNRJ), Museu de Zoologia, Universidade de São Paulo, São Paulo (MZUSP), and Universidad Nacional Autónoma de México, Colección Nacional de Crustáceos, Mexico City (CNCR).Fresh material was fixed in 90% ethanol, preserved in 70% ethanol, and deposited in the INPA collection.All sampling was done according to the permit issued by IBAMA (SISBIO # 44857-1).

Taxonomic and distributional analysis
The taxonomic identification was carried out using the keys, descriptions and illustrations available in Magalhães (1986;2003) and other pertinent publications when necessary (Magalhães and Türkay, 1996a;1996b;1996c;2008;Magalhães and Rodríguez, 2002;Magalhães, 2010), as well as by comparison with identified material in the INPA collection.Suprageneric classification follows Ahyong et al. (2011), and the taxa are organized alphabetically.The distribution maps were generated using the QGIS software, v. 2.14.16 (http://www.qgis.org/pt_BR/site/),following the instructions presented in the second part of the tutorial by Calegari et al. (2016).The geographic coordinates of each record were obtained either from a GPS apparatus during sampling or from data on the labels of the collection lots.When coorditantes were unavaliable, the localities were georeferenced using Google Earth ® software and online gazetteers (GeoNames: http:// www.geonames.org/)and these are noted between parentheses; those coordinates informed between brackets represent the municipality seat, since the locality could not be determined with reasonable accuracy.To generate the maps, the coordinates were all converted to decimal degrees using WGS84 datum (the same datum used in the GPS receiver employed during sampling), according to the calculator provided by the Instituto Nacional de Pesquisas Espaciais (http:// www.dpi.inpe.br/calcula/).

Conventions
Measurements of the crab carapace are in millimeters when they are available; carapace width was measured across the carapace at its widest point, and carapace length was measured along the midline, from the frontal to the posterior margin.Both measurements follow the number and sex of the specimens examined and are given between parenthesis as (carapace width:carapace length); in lots with more than three species, only the measurements of the smallest and the largest specimens are informed.G1 refers to the first male gonopod.Abbreviations used are: "ad." for adult(s); "elev."for elevation; "immat." for immature specimen; "juv." for juvenile(s); "mun." for municipality, "ov." for ovigerous; "trib." for tributary.In 'Material examined' the names of geographic elements were kept as written in Portuguese on the original labels and mean: cachoeira = falls; fazenda = farm; igarapé = small river, stream; lago = lake; PARNA = National Park; praia = beach; PIN = Posto Indígena (Indigenous Health Post); rio = river; serra = mountain range; sítio = small farm; vicinal = side road; vila = village.The names of some landforms, particularly those in the Brazilian territory, were written according to the spelling in Portuguese.

resulTs
The total number of freshwater crab species known in the state of Roraima increased to ten: six Pseudothelphusidae and four Trichodactylidae.Three species are recorded from Roraima for the first time: the pseudothelphusids Fredius beccarii (Coifmann, 1939) and Fredius fittkaui (Bott, 1967), and the trichodactylid Moreirocarcinus laevifrons (Moreira, 1901).The records of F. beccarii also represent the first occurrence of the species in Brazil.
Remarks.The G1s of our specimens (Fig. 1A) are morphological very similar to that of a specimen from the Rupununi River, in Guyana, figured by Rodriguez (1982: 135, fig. 129).
Until now, the species was known from eastern Venezuela and Guyana in the Essequibo River  and Demerara River basins (Rodriguez, 1982).Its occurrence in the Uraricoera River basin indicates that the species can be found in both the northern and southern slopes of the Guyana Shield (Fig. 3).
Remarks.Rodriguez and Pereira (1992) described a new subspecies, F. estevisi siapensis Rodríguez and Pereira, 1992, and distinguished it from the nominal subspecies by the mesial lobe of the G1, presenting "the proximal margin concave and the distal margin regularly arched", giving it a thumb-like appearance, whereas the nominal subspecies has a mesial lobe with proximal and distal margins more or less straight, showing a conical shape.The specimens from Roraima have a G1 with mesial lobe bearing a conical appearance (Fig. 1B, C) and are therefore assigned to the nominal species.Besides, the localities of the Roraima specimens (particularly those of the Miang River) (Fig. 3) are not geographically distant from the localities in the upper Caroni River presented by Rodriguez and Pereira (1992).According to these authors, the nominal form is distributed in the southeastern portion of the state of Bolívar, in the "Gran Sabana" region, mainly in tributaries of the upper Caroni River coming from the slopes of Mount Roraima and draining to the Orinoco River Basin, while F. estevisi siapensis occurs in the Siapa and Cunucunuma Rivers, in the Venezuelan state of Amazonas.
Material examined.BRAZIL, Roraima: 1 male (69.Remarks.The G1s of the Roraima specimens (Fig. 1D) have all the morphological features as described for the G1 of the upper Negro River specimens by Magalhães and Rodríguez (2002): the well-developed, curved cephalic spine with a rounded tip; and an auxiliary lobe situated laterally and shorter than the cephalic lobe.
Brazil and Venezuela (with several records in northern tributaties of the upper Negro River), and in Guyana (Magalhães and Rodríguez, 2002).Unfortunately, the sole record available from Guyana lacks indication of a precise locality (Magalhães and Rodríguez, 2002).The present two records in the state of Roraima (Fig. 4) indicate that the species may have a continuous distribution along the tributaries of the Negro River and Branco River that drain the southern slopes of the Guyana Shield, probably reaching Guyana through the lowland areas along the Tacutu and upper Rupununi Rivers (see also Magalhães and Rodríguez, 2002: 687, fig. 4).
Material examined.BRAZIL, Roraima: 1 male (51.8:31.7),MIRR 43, mun.Amajarí [03°39'20"N 061°25'16"W], serra Tepequém, igarapé da Base, Remarks.The G1s of the specimens studied herein (Fig. 1E, F) have the mesial process wide, with proximal margin distinctly convex in mesio-caudal view, and ending in a bifid tip; cephalic lobe auricular in latero-cephalic view, auxiliary lobe slightly shorter than cephalic one, with caudal border bearing small cephalic spine; field of apical spine well developed, wide, longitudinally situated along the latero-cephalic side.These features highly resemble those of the G1 of the holotype of F. platyacanthus that was figured by Rodriguez and Pereira (1992: 305, fig.3), but some variations can be noticed: the cephalic lobe is a little shorter than the mesial process in the Roraima specimens (even considering the angle from which the illustration was made), and the cephalic spine is a bit smaller than that of the holotype.
Fredius platyacanthus was only known from the male holotype collected in the Icabaru River, a left bank tributary of the upper Caroni River, in the Orinocoan slopes of the Serra Pacaraima.The several records presented above indicate that the species is more widely distributed on the Amazonian slopes of the Guyana Shield.In addition to the records the from the upper and middle course of the Uraricoera River (Fig. 3), Magalhães et al. (2006) recorded the species from the Yanomami village of Balawa-ú, in the state of Amazonas, Brazil, in the headwaters of the Demini River, a left tributary of the Negro River (the authors erroneously noted it as Fredius reflexifrons (Ortmann, 1897) in tab.1).
At the moment, we prefer to consider our material as co-specific with the F. platyacanthus' type, since a more precise assessment of the variability of these G1's features is not possible without examining a more abundant series of specimens from the upper Caroni River area.
Remarks.The gonopods of the specimens from Roraima (Fig. 1G) show a high morphological similarity with that of the holotype as described in Rodríguez and Suárez (1994) and figured in Rodríguez and Campos (1998: 766, fig.2A, B) and no remarkable variation was noticed.
With the present records, F. stenolobus exhibits a somewhat wide distribution that encompasses both the Amazonian and Orinocoan slopes of the Serra Paracarima.On the Venezuelan side, the species is distributed in the contiguous hydrographic basins of the Caura River (Rodríguez and Suárez, 1994;Magalhães and Pereira, 2003) and in the Paragua River (Mora-Day and Blanco-Belmonte, 2008).On the Brazilian side, it seems to be relatively frequent in the more elevated areas (above 750 m) of the upper Uraricoera River and upper Surumu River basins, but also occurring at lower elevations in the middle course of the Branco River after its confluence with the Mucajaí River (lower than 100 m) (Fig. 4).
Remarks.The G1s of the specimens studied herein from the Tacutu River (Fig. 1H) exhibit an overall resemblance with the G1 of the specimen from the Surumu River figured by Magalhães (1986: 629, fig. 8).However, some of the specimens from the Jatapu River basin, in southeastern Roraima have a G1 with more longitudinally straight apical plate, resembling those from the Uatumã River basin (Magalhães, 1986: 629, figs. 9, 10).
The species has a broad distribution that includes several rivers draining both slopes of the Guyana Shield in Brazil (east from the Branco River basin in Roraima to the western portion of the state of Amapá), Guyana, Suriname and French Guiana (Holthuis, 1959;1993;Rodriguez, 1982;Magalhães, 1986;2003;Vieira, 2008; this study).The species is typically found in the upper course of the rivers, usually inhabiting fast flowing, lotic waters with rocky bottoms of rapids and falls (Magalhães, 1986(Magalhães, , 2003;;Holthuis, 1993).
Remarks.This species is easily recognized by the orbicular carapace with lateral margins bearing 7-8 acute teeth and frontal margin bilobed, smooth.The G1 is slender, distally straight and regularly tapering.The apex is narrow and assymetrical, with its mesiodistal border being slightly shorter than the latero-distal border (Fig. 2A).
The species is mainly distributed in the Amazon River and Orinoco River basins in Brazil, Colombia, and Venezuela (Magalhães, 2003;Campos, 2002;2005;2014).In the Brazilian territory, it was only known in the state of Amazonas, in the Negro River basin (Magalhães, 2003).The above records show that it also occurs in the Branco River basin and constitute the first report of the species in the state of Roraima (Fig. 5).
Remarks.With respect to the carapace morphology, P. dentata highly resembles M. laevifrons.However, both species can be readily distinguished by the spined (in P. dentata) or smooth (M.laevifrons) frontal margin.The morphology of the G1 of the specimens studied herein (Fig. 2B) is very similar to the description and illustrations offered by Magalhães and Türkay (2008).
Poppiana dentata has a very wide distribution in northern South America, occurring in the Orinoco River basin and in several coastal river basins from French Guiana to Colombia, as well as on the island of Trinidad and in Nicaragua (Campos, 2005;2014;Magalhães and Türkay, 2008).The distribution of the species in the Amazon River basin seems to be restricted to the Branco River basin, in Roraima (Fig. 5), and the upper Negro River basin, in Amazonas (Magalhães and Türkay, 2008;this study).
Remarks.The G1s of the specimens studied herein (Fig. 2C) are morphological highly similar to the G1 described and illustrated by Magalhães and Türkay (1996b).
Sylviocarcinus pictus is one of the most common and widespread species of Trichodactylidae, occurring in the Amazon River basin and in coastal river basins of northern South America, in Venezuela, Guyana and French Guiana (Magalhães and Türkay, 1996b;Mora-Day et al., 2008).It is also broadly distributed in Roraima (Fig. 6).The species has been recorded from both slopes of the Serra Pacaraima and Serra Acari, with occurences in the Cuyuni River (Mora-Day et al., 2008), Rupununi River (Magalhães and Türkay, 1996b), and upper Essequibo River (Lasso et al., 2008) basins on the Venezuelan and Guyanan sides, and in the Branco River and Jauaperi River, Uatumã River and Trombetas River basins (Magalhães and Türkay, 1996b; this study) on the Brazilian side.However, it does not occur in tributaries of the Orinoco River with headwaters in the Guyana Shield as suggested by its absence in surveys recently done in some river basins of this region (Magalhães and Pereira, 2003;Pereira and García, 2006;Mora-Day and Blanco-Belmonte, 2008;Mora-Day et al., 2008).
The present species has a very broad distribution throughout the entire Amazon River and Orinoco River basins as well as in the coastal river basins of northern South America, but has only recently been reported in the state of Roraima (Castro and Silva, 2013).The above records show that it is also well distributed in the state, occurring mainly in the lowland portion of the Branco River, Jauaperi River and Jatapu River basins (Fig. 7).

disCussion
All ten species of freshwater crabs that are reported herein from the state of Roraima are distributed in the Branco River basin, although K. latifrons, S. pictus and V. serrata can also occur in the Jauaperi River and Jatapu River basins (Figs.3-7).The pseudothelphusid species are usually associated with rivers originating in the mountainouns regions along the border with Venezuela and Guyana in the northwestern (Serra Paraima), northern and northeastern (Serra Pacaraima), and the southeastern portion of the state (Serra Acaraí) (Figs. 3-4), although F. stenolobus can also be found in lower regions along the upper Branco River (Fig. 4).On the other hand, the trichodactylid species are mainly distributed in lower elevation areas along the Branco River and its tributaries, Uraricoera River and Tacutu River (Figs. 5-7), but S. pictus can also occur at higher elevations such as in some tributaries of the Uraricoera River coming from the Serra Pacaraima (Fig. 6).This general altitudinal distribution of representatives of both families in Roraima is in accordance with the zoogeographic division of the Neotropical freshwater decapod fauna mentioned by Rodriguez (1981).
Pseudothelphusidae is slightly richer in species than Trichodactylidae (six vs four) in Roraima, but these numbers may be somewhat underestimated for both families.The more elevated regions along the mountainous areas in the northern and western part of the state are either protected areas or indigenous reserves (EPE, 2010), which makes biological surveys more difficult to conduct in these areas.As a mountainous group, the number of pseudothelphusids might be greater considering that large portions of this territory have not yet been sufficiently explored for decapods.For instance, Kunziana irengis (Pretzmann, 1971) occurs in Guyana along the Ireng River and upper Potaro River basins (Magalhães et al., 2009).The Ireng River marks the border between Brazil and Guyana (it is named Maú River on the Brazilian side) and its presence in the northeastern part of the state of Roraima would not be unlikely.Provided they are properly surveyed, some other regions of the state may potentially increase the number of crab species in Roraima as well.This is the case, for instance, in the Serra da Mocidade, which comprises a mountainous complex with up to 1700 m in elevation, isolated from the Guyana Shield, in the center-western portion of the state.The potential of these isolated mountainous areas in the Amazon region to reveal new pseudothelphusid crabs is high, as indicated by the description of Microthelphusa lipkei Magalhães, 2010 from Serra do Aracá (Magalhães, 2010), also a group of isolated mountains situated in the state of Amazonas close to the border with Roraima, a little to the west of the Serra da Mocidade.
The number of trichodactylid species from Roraima may also be somewhat underestimated.Extensive lowland areas of the lower Branco River are still poorly explored for freshwater crabs (Figs.5-7).Magalhães and Pereira (2007) listed eight species as occurring in the Negro River, of which the Branco River is the main tributary, and four of them (M.laevifrons, P. dentata, S. pictus, V. serrata) are also recorded from Roraima henceforth.It would not be unexpected that at least some of the others could also be found in the lower Branco River if more intensive surveys of the area were carried out.
As expected, the species richness of freshwater crabs of the Branco River is lower than that of Negro River, which has a much larger drainage surface and is better surveyed, especially in its lower course.Magalhães and Pereira (2007) listed 10 pseudothelphusids and eight trichodactylids from the Negro River basin, while six and four, respectively, are herein reported from the Branco River basin.In total, the species richness of crab of the Xingu River basin is higher than that of the Branco River basin (13 vs. 10 species, respectively), but Branco River basin crab fauna is a little richer in pseudothelphusids than Xingu River basin fauna (six vs. five species, respectively).However, Xingu River has more trichodactylids than the Branco River (eight vs four, respectively) (Magalhães et al., unpubl. data).When comparing the total number of species compiled by Magalhães and Pereira (2007) to the sub-basins of the other northern tributaries of the Amazon basin draining the Guyana Shield, the Branco River basin shows higher species richness (Uatumã River: six species; Trombetas River: seven species; Jari River: four species).However, such comparisons should be made with great care as these river basins are very poorly surveyed for decapods.2011. Subphylum Crustacea Brünnich, 1772. p. 165-191.

aCKnowledgeMenTs
The authors dedicate this contribution to Prof. Dr. Ludwig Buckup (Universidade Federal do Rio Grande do Sul, retired) in recognition of his distinguished career as a zoologist and particularly for his achievements that have greatly contributed to the development of the Brazilian Carcinology both scientifically and in the formation of young scientists.PMC also thanks CNPq for the Research Project 35.0413/2005-3-DCR/ CNPq/FEMACT; FZ and PMC thank the Acordo de Cooperação Técnico-Científica 01280.000370/2011 between INPA, UERR, and IACTI-RR that allowed the accomplishment of the present work.CM thanks the Conselho Nacional de Desenvolvimento Científico e Tecnológico -CNPq for an ongoing Research Scholarship 304736/2015-5, as well as Victor Py-Daniel, Ulysses C. Barbosa, and Jansen Zuanon for making the specimens from their collections in Roraima available to this study.The authors are grateful to Douglas Bastos and Priscila Madoka Ito for their kind help with the QGIS software, Colleen M. Flanagan for reviewing the English wording, and two anonymous reviewers for their comments.referenCesAhyong, S.T.; Lowry, J.K.; Alonso, M.; Bamber, R.N.; Boxshall, G.A.; Castro, P.; Gerken, S.; Karaman, G.S.; Goy, J.W.; Jones, D.S.; Meland, K.; Christopher, D. Rogers and Svavarsson, J.